A5012459556- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Neural dynamics and brain function
- CCD and CMOS Imaging Sensors
- Photoreceptor and optogenetics research
- Transition Metal Oxide Nanomaterials
- Semiconductor materials and devices
- Neural Networks and Applications
- Magnetic properties of thin films
- Modular Robots and Swarm Intelligence
- Antenna Design and Optimization
- Neural Networks and Reservoir Computing
- Advanced biosensing and bioanalysis techniques
- Wireless Communication Networks Research
- Conducting polymers and applications
- Satellite Communication Systems
CEA LETI
2018-2024
Commissariat à l'Énergie Atomique et aux Énergies Alternatives
2018-2024
Université Grenoble Alpes
2018-2024
National University of Singapore
2022
CEA Grenoble
2018-2019
Institut polytechnique de Grenoble
2018-2019
Abstract Phase change memory can provide a remarkable artificial synapse for neuromorphic systems, as it features excellent reliability and be used an analog memory. However, this approach is complicated by the fact that crystallization amorphization differ radically: realized in very gradual manner, similarly to synaptic potentiation, while process tends abrupt, unlike depression. Addressing non‐biorealism of requires system‐level solutions have considerable energy cost or limit generality...
We demonstrate, for the first time, 3D monolithically integrated multiple 1T1R Resistive RAM (RRAM) structure storing up to 3.17 bits per RRAM. study, using a 4 kb array, impact of conductance relaxation after Multi-Level Cell (MLC) programming. show that traditional storage applications may be limited 2 RRAM due overlap between ranges relaxation. On other hand, our study concludes effect is negligible Neural Network (NN), allowing use nine distinct levels (equivalent bits) with minimal...
Recurrent neural networks are currently subject to intensive research efforts solve temporal computing problems. Neuromorphic processors (NPs), composed of networked neuron and synapse circuit models, natively compute in time offer an ultralow power solution particularly suited emerging edge-computing applications (wearable medical devices, for example). The most significant roadblock addressing useful problems with neuromorphic hardware is the difficulty maintaining healthy network dynamics...
Spiking neural networks (SNNs) are a computational tool in which the information is coded into spikes, as some parts of brain, differently from conventional (NNs) that compute over real-numbers. Therefore, SNNs can implement intelligent extraction real-time at edge data acquisition and correspond to complementary solution NNs working for cloud-computing. Both NN classes face hardware constraints due limited computing parallelism separation logic memory. Emerging memory devices, like...
Resistive switching memories (RRAMs) have attracted wide interest as adaptive synaptic elements in artificial bio-inspired spiking neural networks (SNNs). These devices suffer from high cycle-to-cycle and cell-to-cell conductance variability, which is usually considered a big challenge. However, biological synapses are noisy the brain seems some situations to benefit noise. It has been predicted that RRAM-based SNNs intrinsically robust variability. Here, we investigate this robustness based...
Resistive Random Access Memories (RRAMs) are a promising solution to implement Ternary Content Addressable (TCAMs) that more area- and energy-efficient with respect Static Memory (SRAM)-based TCAMs. However, RRAM-based TCAMs limited in the number of bits per word due low ratio between resistances high resistance states (HRS/LRS) variability RRAM. Such limitation on length hinders parallel search very large data for data-intensive applications. To overcome this issue, first time, we propose...
Resistive Memory (RRAM)-based Ternary Content Addressable Memories (TCAMs) were developed to reduce cell area, search energy and standby power consumption beyond what can be achieved by SRAM-based TCAMs. In previous works, RRAM-based TCAMs have already been fabricated, but the impact of RRAM reliability on TCAM performance has never proven until now. this work, we fabricated extensively tested a circuit. We show that trade-off exists between latency in terms match/mismatch detection...
This paper provides an overview of the challenges faced by hardware implemented Spiking Neural Networks, from device to circuit design, reliability and test. We present a comprehensive description state-of-the-art neuromorphic architectures inspired brain computation, with special emphasis on Networks (SNNs), together emerging technologies that have enabled such systems, namely Phase Change Metal Oxide Resistive Memories. Finally, we discuss main implementations SNNs, their post-fabrication...
Resistive Random Access Memory (RRAM)-based artificial Neural Networks (NNs) have been shown to be intrinsically robust RRAM variability but no study has done clearly explain and quantify this robustness. In paper, we fully characterize a 4kbit array under different programming conditions. The impact of the electrical characteristics (resistance variability, memory window, endurance performance) on detection rate NN designed for object tracking trained with stochastic Spike-Timing Dependent...
Brain-inspired computing systems are attracting considerable attention due to their high energy efficiency with respect conventional when applied problems in artificial intelligence, sensing and robotics. This is attributed the spike-based computational mechanism architectural co-localization of processing memory. To realize this neuron circuits must be integrated further modelling synapses. Synapses require exhibit plasticity, that modulation efficacy, support online learning algorithms,...
The memtransistor (MT) based on 2-D materials is considered a promising candidate to extend Moore's law, thanks its scalability. As hybrid integration of memristor and transistor, MT combines the electrical properties memristor-based one-transistor one-resistor (1T-1R) structures (i.e., nonvolatile resistive switching gate selection) into single compact device. In this work, we evaluate technology from an application perspective explore use in ternary content-addressable memory (TCAM)...